Diagnosing Catheter- Related Microbaemia
ABSTRACT & COMMENTARY
Synopsis: A meta-analysis showed that a single quantitative blood culture withdrawn from the catheter was the single most cost-effective means of diagnosing catheter-related microbaemia.
Source: Siegman-Igra, et al. Diagnosis of vascular catheter related bloodstream infection: A meta analysis. J Clin Microbiol 1997;35:928-936.
The diagnosis of catheter-related blood- stream infections (or microbaemia, which I prefer), as distinct from other catheter-related infections such as those of the exit site, is known to be difficult while the device is in place. Though diagnosis is easier after removal, the information is seldom of use to the clinician. The incidence of these infections continues to rise in line with the ever-increasing number of access devices being employed. Needless to say, there are many different methods used for attempting to diagnose infection and little clarity on their relative accuracies and cost-effectiveness. Siegman-Igra et al, therefore, undertook a meta-analysis of 22 published studies out of the 156 they identified for possible inclusion to address these issues.
Catheter-related microbaemia was defined in most studies as the presence of clinical features of microbaemia, recovery of the same organism from peripheral blood, and either the catheter segment or blood drawn for culture via the device (lumen blood culture) in the absence of any other possible source of the infection. The sensitivity and specificity of the test methods were estimated for qualitative, semi-quantitative, and quantitative catheter segment culture, qualitative and quantitative lumen blood culture (so-called unpaired blood culture), and quantitative paired lumen and peripheral blood cultures.
The quantitative lumen blood culture was the most specific test, whereas the qualitative segment culture was the most sensitive. The receiver operating characteristic (ROC) curves resulting from segment cultures indicated that these resulted in higher accuracy, and they also tended to be more sensitive and less specific than were blood cultures. However, despite their relatively lower unit costs, they were more expensive in terms of cost per accurate result. Quantitative catheter segment culture was the only method to yield a sensitivity and specificity exceeding 90%, but a quantitative lumen blood culture was the single most cost-effective method even though its sensitivity was low.
COMMENT BY J. PETER DONNELLY, PhD
Given the increasing number of catheters in use and the attendant problems associated with them, it would be surprising if this report were to remain largely ignored. Moreover, such a meta-analysis ought to provide at least some useful information for helping choose a method for diagnosing catheter-related microbaemia. However, I have some reservations. In the first place, the numbers of patients with catheter-related microbaema were small, ranging from four to 36 with an average of 11 patients per study. Second, while the thresholds used for defining a positive test result for a quantitative catheter segment culture were 100-1000 cfu, those for qualitative blood cultures ranged from any growth to more than 100 cfu, and those for a lumen blood culture from 15 to less than 100 cfu (i.e., a 6-fold difference). Either the ratio of paired peripheral blood and lumen cultures was 3:1, 4:1, or 5:1, or there had to be a difference of 30 cfu between the lumen and peripheral blood cultures.
Aside from these difficulties, the authors of necessity used the average sensitivity (and specificity) for each test studied which were broadly representative of the data, but because of the small sample size any center applying the same tests would most probably not get the same results. The ROC partially compensated for this by showing a trend in favor of catheter-segment cultures in terms of producing the lowest proportion of false positive results, but, again, this might not be borne out of practice. However, my central concern is that the results do seem out of kilter with the everyday hustle and bustle of the ward.
Once in place, clinicians are, with good reason, reluctant to pull catheters unless they are convinced of refractory infection, an infection that is more commonly understood to mean cutaneous infection around the device rather than catheter-related microbaema. Moreover, most clinicians maintain a sanguine disposition to Staphylococcus epidermidis, which is by far the most common isolate involved, and would be more inclined to attempt treatment or even do nothing. However, microbaemia with other organisms, such as the Gram-negative bacilli and yeasts, would prompt them to seriously consider removing the line. The authors did not make any distinction between these two entirely different entities. Also, the authors included as their definition of catheter-related bacteremia "any clinical features of microbaemia in the absence of another possible source of the infection," leaving me to wonder what they might be. Fever, I suspect, and little else, in which case treatment would probably be given empirically whether or not the blood cultures yield growth, and would most likely be with teicoplanin or vancomycin. Treatment would then only be altered should any organism recovered from peripheral blood and/or blood drawn for culture via the device prove to be other than a coagulase-negative Staphylococcus. A culture of the catheter segment would play little or no role in this and would only serve to confirm after the event.
It would be nice if the paper’s most appealing conclusion that a single lumen quantitative blood culture would be the most cost-effective for obtaining an accurate result were to prove correct, not least because the laboratory is almost guaranteed a sample. Moreover, this would provide a meaningful point of contact for the laboratory and clinician, who would have some chance of altering management. But, are clinicians to rush into only asking for this and abandon catheter segment cultures altogether as these only add to the costs? These pieces of plastic are fascinating to a microbiologist not least because his or her ingenuity is tested to the limit in trying to tease out the most recalcitrant microorganism by rolling, flushing, scraping, shaking, or sonicating the tip. Moreover, they provide wonderful electron micrographs. However, a perusal over the myriad methodological articles shows how much a matter of preference this whole business is and how little effect the results have on clinical decision making.
Nevertheless, I can see a lot of unhappy microbiologists if they were to be denied yet another of life’s simple pleasures. More seriously, there would also be no gold standard by which to validate diagnosis. Whatever my quibbles, Siegman-Igra et al have done us a service by providing a useful pointer to future formal prospective studies and have therefore achieved the primary goal of meta-anlysis. Catheter-related infections are iatrogenic by nature and can prove expensive; therefore, any future study ought to attempt to estimate the costs saved by timely diagnosis, at least in terms of adminstering less antimicrobial therapy.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content